Visual perceptual learning (VPL) refers to a long-term enhancement of visual task performance. VPL is an important tool to improve and/or restore damaged/ degraded vision. VPL is usually specific for the trained stimulus. Although specificity in VPL has given great insights into the mechanisms of VPL, it is also regarded as a serious limitation to clinical applications that require generalized effects. In this proposal, we raise the possibility that VPL of natural and everyday scenes is less specific for the trained stimulus than VPL of a simple and unnatural stimulus. In our daily life, the visual system actively interact with or is exposed to complex natural or everyday scenes composed of categories and other characteristics. Thus, in Aims 1 and 2, we will first test how categorization influences the degree of specificity in VPL using simple but categorized stimuli. In Aim 3, we will examine whether and how transfer of VPL occurs as a result of being exposed to NS images. Specific Aim 1 will examine the effect of category learning on transfer of VPL (category-induced transfer of VPL, CIT-VPL). The following hypotheses will be tested by psychophysics and brain imaging. Hypothesis (H) 1-1a: CIT-VPL occurs only within the same category as the trained. H1-1b: CIT-VPL is greater with orientations closer to the category boundary. H1-2: CIT-VPL results from performance increases on untrained orientations within the trained category (a) or performance decreases on untrained orientations an untrained category (b). H1-3: CIT-VPL occurs in a location-specific manner . H1-4: Early visual areas play a role in inducing CIT-VPL. Specific Aim 2 will examine whether and how much CIT of VPL of orientation is generalized to CIT of VPL of motion direction. H2-1 to H2-4 will be identical to H1-1 to H1-4, except that CIT of VPL of motion direction, instead of orientation, will be tested. Specific Aim 3 will examine how orientation specificity of VPL is influenced by statistics and categories of natural scenes. We will the following hypotheses by psychophysics. H3-1: VPL occurs as a result of exposure to NSs. This hypothesis will be tested by VPL of orientations after subjects are exposed to task-irrelevant natural scenes with a certain mean orientation spectrums. H3-2: VPL of NSs as task-irrelevant is less location specific than VPL of unnatural stimuli. We will test this hypothesis by examining whether VPL of natural scenes on one side of the visual field transfers to the other side. H3-3: VPL of a task-relevant stimulus is influenced by the category of a background NS as task-relevant.